Self-checkout went in. Face-to-face service did not go down.
Four stores. Measured again one week in, and again three months in.
From problem to result
- Problem
- Queueing and paying at the counter took a long time
- Applied and measured
- The worry was that self-checkout would cut the time staff and shoppers spend together. After rolling it out to four stores we measured two points, one week in and three months in, with saai insight, separating checkout duration from contact outside the checkout area.
- Measured result
- Checkout duration fell sharply, while staff and shopper contact at the fixtures held or rose, store by store
saai insight
47% to 91% down
Checkout duration (three months in · range across stores)
- Checkout duration (smallest gain)-47%
- Checkout duration (largest gain)-91%
- Fixture that worked · visits+60%
- Fixture that worked · dwell+120%
Checkout duration is the three-month re-measurement; only the two ends of the range across four stores are shown. Colour marks direction of change only. The bar on the home card uses the most conservative point.
Held or rose, store by store
Staff and shopper contact outside the checkout area
Visits 1.6x · dwell 2.2x after the move
Where the fixture move worked
Moved to a more visible spot, yet visits and dwell both fell
Where the fixture move missed
Evidence and scope
Adoption stage
- Pilot
- Deployed
- Scaled
How this was verified
Measurements from one week in and three months in were compared for the same stores and the same time bands. Branch names are not disclosed.
A real deployment across four stores of a major Korean health and beauty chain. Branch names and absolute figures are banded and reconstructed to protect the client; the rates of change are as measured.